hohn/codeql
1
0
Fork 0
mirror of https://github.com/github/codeql.git synced 2025-12-17 01:03:14 +01:00
Code Issues Packages Projects Releases Wiki Activity

C#: Fix glob pattern processing: allow **/ to match empty string

Browse Source
This commit is contained in:
Tamas Vajk
2024-07-03 08:09:30 +02:00
parent c7ad0ad406
commit b36db5ad11
3 changed files with 54 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
csharp/extractor/Semmle.Extraction.Tests/FilePathFilter.cs
View File

@@ -216,9 +216,9 @@ namespace Semmle.Extraction.Tests
var expectedRegexMessages = new[]
{
"Filtering in files matching '^c/.*/i\\.[^/]*.*'. Original glob filter: 'include:c/**/i.*'",
"Filtering in files matching '^c/d/.*/[^/]*\\.cs.*'. Original glob filter: 'include:c/d/**/*.cs'",
"Filtering out files matching '^.*/z/i\\.cs.*'. Original glob filter: 'exclude:**/z/i.cs'"
"Filtering in files matching '^c/(.*/|)i\\.[^/]*.*'. Original glob filter: 'include:c/**/i.*'",
"Filtering in files matching '^c/d/(.*/|)[^/]*\\.cs.*'. Original glob filter: 'include:c/d/**/*.cs'",
"Filtering out files matching '^(.*/|)z/i\\.cs.*'. Original glob filter: 'exclude:**/z/i.cs'"
};
Assert.Equivalent(expectedRegexMessages, logger.Messages, strict: false);
}
@@ -244,10 +244,41 @@ namespace Semmle.Extraction.Tests
var expectedRegexMessages = new[]
{
"Filtering in files matching '^.*/i\\.[^/]*.*'. Original glob filter: 'include:**/i.*'",
"Filtering out files matching '^.*/z/i\\.cs.*'. Original glob filter: 'exclude:**/z/i.cs'"
"Filtering in files matching '^(.*/|)i\\.[^/]*.*'. Original glob filter: 'include:**/i.*'",
"Filtering out files matching '^(.*/|)z/i\\.cs.*'. Original glob filter: 'exclude:**/z/i.cs'"
};
Assert.Equivalent(expectedRegexMessages, logger.Messages, strict: false);
}
[Fact]
public void TestFiltersWithIncludeExcludeComplexPatternsRelativeRoot()
{
(var testSubject, var logger, var files) = TestSetup();
// 'c' is the start of the relative path so we want to ensure the `**/` glob can match start
Environment.SetEnvironmentVariable("LGTM_INDEX_FILTERS", """
include:**/c/**/i.*
exclude:**/c/**/z/i.cs
exclude:**/**/c/**/z/i.cs
""");
var filtered = testSubject.Filter(files);
var expected = GetExpected(
[
"/a/b/c/x/y/i.cs",
]);
AssertFileInfoEquivalence(expected, filtered);
var expectedRegexMessages = new[]
{
"Filtering in files matching '^(.*/|)c/(.*/|)i\\.[^/]*.*'. Original glob filter: 'include:**/c/**/i.*'",
"Filtering out files matching '^(.*/|)c/(.*/|)z/i\\.cs.*'. Original glob filter: 'exclude:**/c/**/z/i.cs'",
"Filtering out files matching '^(.*/|)(.*/|)c/(.*/|)z/i\\.cs.*'. Original glob filter: 'exclude:**/**/c/**/z/i.cs'"
};
Assert.Equivalent(expectedRegexMessages, logger.Messages, strict: false);
}
}
}

6
csharp/extractor/Semmle.Extraction.Tests/FilePattern.cs
View File

@@ -10,13 +10,15 @@ namespace Semmle.Extraction.Tests
var fp = new FilePattern("/hadoop*");
Assert.Equal("^hadoop[^/]*.*", fp.RegexPattern);
fp = new FilePattern("**/org/apache/hadoop");
Assert.Equal("^.*/org/apache/hadoop.*", fp.RegexPattern);
Assert.Equal("^(.*/|)org/apache/hadoop.*", fp.RegexPattern);
fp = new FilePattern("hadoop-common/**/test// ");
Assert.Equal("^hadoop-common/.*/test(?<doubleslash>/).*", fp.RegexPattern);
Assert.Equal("^hadoop-common/(.*/|)test(?<doubleslash>/).*", fp.RegexPattern);
fp = new FilePattern(@"-C:\agent\root\asdf//");
Assert.Equal("^C:/agent/root/asdf(?<doubleslash>/).*", fp.RegexPattern);
fp = new FilePattern(@"-C:\agent+\[root]\asdf//");
Assert.Equal(@"^C:/agent\+/\[root]/asdf(?<doubleslash>/).*", fp.RegexPattern);
fp = new FilePattern(@"**/**/abc/**/def/**");
Assert.Equal(@"^(.*/|)(.*/|)abc/(.*/|)def/.*", fp.RegexPattern);
}
[Fact]

16
csharp/extractor/Semmle.Extraction/FilePattern.cs
View File

@@ -76,8 +76,20 @@ namespace Semmle.Extraction
throw new InvalidFilePatternException(pattern, "'**' preceeded by non-`/` character.");
if (HasCharAt(i + 2, c => c != '/'))
throw new InvalidFilePatternException(pattern, "'**' succeeded by non-`/` character");
sb.Append(".*");
i += 2;
if (i + 2 < pattern.Length)
{
// Processing .../**/...
sb.Append("(.*/|)");
i += 3;
}
else
{
// Processing .../**
sb.Append(".*");
// There's no need to add another .* to the end outside the loop, we can return early.
return sb;
}
}
else
{